health check: daily eval 2026-07-30 [automated]

.eval_results/health-check-embedding-multilingual-2026-07-30.yaml

@@ -0,0 +1,99 @@
+# Health Check: Nanthasit/sakthai-embedding-multilingual
+# Generated: 2026-07-30T12:00:00+00:00
+
+model:
+  id: Nanthasit/sakthai-embedding-multilingual
+  scanned_at: '2026-07-30T12:00:00+00:00'
+  pipeline_tag: sentence-similarity
+  library_name: sentence-transformers
+  private: false
+popularity:
+  downloads: 362
+  likes: 0
+  age_days: 5
+  velocity_daily: 72.4
+  author_rank: 6
+  author_velocity_rank: 1
+  author_share_pct: 6.2
+assessment:
+  status: fair
+  score: 55
+  breakdown:
+    popularity: 13
+    download_momentum: 89
+    benchmark_coverage: 0
+    card_quality: 75
+    repo_hygiene: 100
+  strengths:
+  - 'Velocity leader: 72.4/day (rank 1/18)'
+  - 'Clean repo: 0 dev artifacts'
+  - Full model.safetensors (471 MB), real weights
+  concerns:
+  - No MTEB/BEIR benchmarks in model card
+  - 0 likes after 5 days — low community signal
+  - 'Embedding model niche: slower organic growth vs text-gen models'
+  recommendations:
+  - Publish MTEB benchmark results (STSBenchmark, multilingual retrieval)
+  - Add sentence-transformers usage examples to README
+  - Create demo Space for cross-lingual semantic search
+architecture:
+  model_type: bert
+  architecture: BertModel
+  embedding_dimension: 384
+  pooling_mode: mean
+  similarity_fn: cosine
+  hidden_size: 384
+  num_hidden_layers: 12
+  num_attention_heads: 12
+  intermediate_size: 1536
+  max_position_embeddings: 512
+  vocab_size: 250037
+  parameters_estimate: ~118M (BERT-base class)
+  dtype: float32 (no quantization)
+  prompts:
+    query: ''
+    document: ''
+file_inventory:
+  model_safetensors: 470.6 MB
+  tokenizer_json: 17.1 MB
+  total_siblings: 14
+  non_hidden_files: 9
+  dev_artifact_count: 0
+  has_1_Pooling_config: true
+  has_sentence_transformers_config: true
+card_metadata:
+  license: mit
+  language:
+  - multilingual
+  tags_count: 14
+  tags_sample:
+  - sentence-transformers
+  - embeddings
+  - sentence-similarity
+  - feature-extraction
+  - cross-lingual
+  - semantic-search
+  - retrieval
+  - rag
+  datasets_count: 7
+  datasets:
+  - Nanthasit/sakthai-combined-v6
+  - Nanthasit/sakthai-combined-v7
+  - Nanthasit/sakthai-kaggle-notebooks
+  - Nanthasit/sakthai-irrelevance-supplement
+  - Nanthasit/food-penguin-v1
+  - Nanthasit/SimpleToolCalling
+  - Nanthasit/sakthai-bench-v1
+  has_model_index: false
+eval_metadata:
+  source_endpoints:
+  - /api/models/Nanthasit/sakthai-embedding-multilingual?expand[]=cardData
+  - /api/models/Nanthasit/sakthai-embedding-multilingual?blobs=true&expand[]=siblings
+  - config.json, 1_Pooling/config.json, config_sentence_transformers.json
+  - /api/models?author=Nanthasit&limit=50
+  scanned_at: '2026-07-30T12:00:00+00:00'
+  scoring_adjustments:
+  - base_model deduction skipped (sentence-similarity)
+  - model-index deduction softened from -30 to -10 (embedding model)
+  - description deduction skipped (README >= 2 KB)
+  changes_since_previous: First automated health check